© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 49152
    9. 8-RES_FS49152-JB
    11. 26506
    Prev
    Next
    INDEX
    Explanations

    references to institutions and organizations in a political context

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    87
    +0.16
    0.8%
    481
    +0.10
    0.5%
    454
    +0.09
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    401
    +0.16
    0.07
    288
    +0.10
    0.07
    224
    +0.09
    0.06
    Negative Logits
     sqor
    -0.68
    foundland
    -0.65
    MpServer
    -0.62
    ranging
    -0.60
    iannopoulos
    -0.59
    range
    -0.59
    eworthy
    -0.56
    erville
    -0.56
    ãĤ¨ãĥ«
    -0.55
    rising
    -0.55
    POSITIVE LOGITS
     intervened
    1.08
     deems
    1.03
     decides
    0.96
     hadn
    0.92
     interfered
    0.92
     couldn
    0.88
     refuses
    0.86
     approves
    0.85
     considers
    0.85
     interven
    0.84
    Activations Density 0.337%

    No Known Activations